[SOLVED] Forecasting using historical seasonally effected )non-linear) data

Forecasting using historical seasonally effected data

Hello,

I organize my thoughts in numbers so it will be easier to understand.

[1] Say I have this data from January 2009 to December 2009:

1-4-7-10-13-14-15-16-18-20-22-25

As you can see, the growth is not linear.

[2] So then, I start with this data point for Janaury 2010.

26-a-b-c-d-f-g-h-i-j-k-l

[3] I now want to forecast that data point for the rest of 2009...using the growth in [1]. SO i guess that's an easy task right? Take the percentage change from Jan 08 and Feb 08...and apply that to the 26...and so on. (am i correct here at least).

[4] But lets say we have a GOAL. So, this become a forecast/plan. In December 2010, i want to grow double the amount of that in December 09. So 25 x 2 = 50. Now, the "known" data is like this:

26-a-b-c-d-f-g-h-i-j-k-50

So how do i do this now?

Any help is greatly appreciated.

It actually wasn't one year, it's an average of several years. And that's beside the point really.

Hey shg,

Thanks for the file and the insight. Greatly appreciated.

I'll tell you at the onset that the data above, i made it up just as an example.

Let me let you know exactly whats going on.

I work at a news company...ie; traffic stats is the importance.

Each month, bonus for editors is determined if they meet the target page view goal.

How do we get the goal for each month? It's hasn't been done in the best ways before so I will not imitate those methods. But we know that by the time we get to December 2010...we want it double of December 2009.

I also have monthly Total U.S. Internet Audience stats from comScore (how many people used the interest in the U.S. in that month...this is what I will be using to determine seasonal fluctuations.

What I did with the total US stats....i got the m/m $ for the past 3 years, and averaged every month. This would be the 2009, column B and C in the file you sent...as in if i were to put the real data.

If you could let me know how you got te 6.59 figure that would be greatly appreciated.

Just realized this isn't going to work out.

i cant model the growth according to the growth of the internet audience....because i have a goal (double december)...so the first month is a 45% increase in traffic.

What i want. Start from 25, end at 50....what do you suggest i do? Forget about my model. What do you suggest i do.....to also reflect that people go online less in the summer, more in the winter.

thank you very much, i really appreciate it. I ended up using your model.

Basically it is a constant growth %...and each month the seasonal adjustment is added to it.

Excellent and thanks alot.